What is MCP-Server?
MCP-Server is a basic math server built using the FastMCP framework that enables basic arithmetic operations like addition, subtraction, multiplication, and division within MCP-compatible environments.
How to use MCP-Server?
To use MCP-Server, install the required dependencies, run the server using the uv
command, and configure your MCP-compatible interface (like Claude Desktop) to connect to the server.
Key features of MCP-Server?
- Supports basic arithmetic operations: addition, subtraction, multiplication, and division.
- Provides a help resource for tool usage.
- Can be integrated with any MCP-compatible interface for seamless operation.
Use cases of MCP-Server?
- Performing basic math calculations in real-time.
- Integrating with language models for enhanced computational capabilities.
- Serving as a backend for educational tools that require arithmetic operations.
FAQ from MCP-Server?
- What operations can MCP-Server perform?
MCP-Server can perform addition, subtraction, multiplication, and division.
- Is MCP-Server free to use?
Yes! MCP-Server is open-source and free to use.
- What are the prerequisites for running MCP-Server?
You need Python 3.9+, Claude Desktop, and the
uv
runtime installed.
Basic Math MCP Server
A simple yet powerful MCP server built using the FastMCP framework that enables basic math operations like addition, subtraction, multiplication, and division — usable right inside Claude Desktop or any MCP-compatible environment.
What is MCP?
MCP (Model Context Protocol) is an open standard developed by Anthropic that allows external tools, APIs, or scripts to be exposed to large language models (LLMs) like Claude in a structured and interactive way.
With MCP, developers can:
- Build custom tools accessible by language models.
- Extend capabilities of Claude with Python scripts or microservices.
- Share structured data or custom responses via
@tool()
and@resource()
decorators.
What is a Custom MCP Server?
A Custom MCP Server is a Python-based service that implements the MCP specification using the mcp-server
library. These servers:
- Define callable tools (
@tool
) for computation or automation. - Host resources (
@resource
) like documentation or reference data. - Run locally or remotely, responding to MCP clients such as Claude Desktop.
MCP servers enable low-latency, secure, and contextual integration between your own code and the capabilities of LLMs.
What We’ve Built
This project is a minimal but complete MCP server offering basic arithmetic capabilities. Tools exposed:
Tools:
add(a, b)
– Returns the sum of two numbers.subtract(a, b)
– Returns the result of subtractingb
froma
.multiply(a, b)
– Returns the product ofa
andb
.divide(a, b)
– Returns the quotient ofa
divided byb
. Handles division by zero.
Resources:
calculator://help
– A Markdown-formatted help file that explains usage of all tools.
Prerequisites
Before getting started, make sure you have the following installed:
- Python 3.9+
- Claude Desktop (or any MCP-compatible interface)
- uv runtime (for fast, isolated Python execution)
Installation Steps
uv
(required to run MCP servers)
Step 1: Install macOS/Linux:
curl -Ls https://astro.build/install/uv | bash
This will install uv
to ~/.local/bin/uv
Windows (PowerShell):
irm https://astro.build/install/uv.ps1 | iex
You may need to restart your terminal or add uv
to PATH.
Step 2: Install the MCP Server
Use uv
to install mcp-server
and run your server:
uv pip install mcp-server
To run the server manually:
uv run mcp-server
Integration with Claude Desktop
Once installed, you need to configure Claude to discover this MCP server.
claude_desktop_config.json
Locate This file holds MCP server configurations for Claude Desktop. Add or update as below.
If You Don’t Have Any MCP Servers:
Paste this entire block into your config:
{
"mcpServers": {
"mcp-server": {
"command": "uv",
"args": [
"run",
"mcp-server"
]
}
}
}
If You Already Have MCP Servers:
Just add the following entry to the "mcpServers" object:
"mcp-server": {
"command": "uv",
"args": [
"run",
"mcp-server"
]
}
Make sure your JSON stays valid (e.g., commas between entries).
Usage Examples
This image illustrates how the Basic Math MCP Server works in real-time with Claude Desktop.
Tool Help Resource
You can also use the built-in help:
resource: calculator://help
It returns markdown-formatted instructions for all tools.
Useful Links
- PyPI: MCP-Server
- Claude + MCP Docs: Official Anthropic Guide
- GitHub UV Runtime: https://github.com/astral-sh/uv